Wall Effects in Photochemically Induced Chain Reactions

that even in vessels having diameters of several centimeters a very large fraction of photochemically produced chains may terminate on the wall unless...
0 downloads 0 Views 607KB Size
July, 1951

WALL

EFFECTS IN

PHOTOCHEMICALLY INDUCED CHAIN

0.034 mg./ml. per degree over the range from 15 to 50°, and a solubility of 1.35 mg./ml. a t 20'. A tenfold reduction in the solubility can be obtained by the addition of an equal volume of ethanol to the water solution. Moreover the concentration of the thiosulfate in the supernatant can be decreased further by the presence of excess (Niena)++. The fact that less than 0.1 mg./ml. of the thiosulfate ion remains in solution makes Ni eno(NO& a very suitable reagent for quantitative precipitation of thi~sulfate.~However, in the presence of sulfate, some Ni en8SOcis coprecipitated by alcohol from water solution, even though it is considerably more soluble than the thiosulfate. This precludes determination of the Ni en8203 spectrophotometrically because of the absorption of the sulfate salt in ~ l t r a v i o l e t but , ~ it does not

[CONTRIBUTION FROM THE

3039

interfere with the determination of the thiosulfate iodometrically.lo No attempt has been made in this study to ascertain the solubility of thiosulfate salts containing less than three molecules of ethylenediamine. The presence of additional ethylenediamine in small amounts was found to have no effect on the solubility in water a t 20'. This seems to indicate that lower complexes were not being formed to influence the solubility data. Alcohol precipitation did not affect the composition of the triethylenediamin&ckel thiosulfate, since no change in the O.D. or iodine equivalent was observed. (IO) K.Arai, F. L. Aldrich and J. H.Gast, Federation Pvoc., 9, 146 (1950).

HOUSTON, TEXAS

(9) J. H. Gast aud F. L.Aldrich, unpublished data.

REACTIONS

DEPARTMENT OF CHEMISTRY,

RECEIVED DECEMBER 20, 1Y6U

COLUMBIA UNIVERSITY ]

Wall Effects in Photochemically Induced Chain Reactions BY RICHARD M. NOYES A mathematical solution has been developed for the situation in which photochemically produced atoms or radicals may combine homogeneously at a rate which is second order in the reactants, and simultaneously may combine heterogeneously on the wall by a iirst order mechanism. The treatment suggests that the walls may serve to remove those atoms or radicals located in a shell of thickness 0.9(DB/Qk)'/dwhere D is the diffusion coefficient, Q is the rate of production of chains and k is the rate constant for homogeneous termination. If the volume of this shell is significant compared t o the total volume of the system, the walls may be responsible for terminating an important fraction of the chains which are formed. It appears that even in vessels having diameters of several centimeters a very large fraction of photochemically produced chains may terminate on the wall unless the absorption of radiation is much greater than 10'1 quanta/ml. sec. and/or the pressure is considerably in excess of atmospheric. The treatment also shows the applicability of kinetic data to determine the probability that a n atom or radical striking the wall of the vessel will undergo recombination.

A great many gas-phase photochemical reactions involve chain mechanisms in which the active intermediates are atoms or radicals produced by the primary photochemical process. The chains may be terminated either homogeneously in the gas phase or heterogeneously on the yalls of the vessel. The homogeneous termination reaction is usually second order in the active intermediate. The heterogeneous termination reaction is usually first order in the active intermediate except a t elevated temperatures.' As a result of the heterogeneous termination a gradient will be established in the concentration of active intermediate. The exact mathematical treatment of this situation is complex, and experimenters have frequently claimed that wall effects could be ignored but have failed to provide adequate justification. Rashevsky2 has integrated the diffusion equation for a number of biological systems; however, he does not consider the combined generation and destruction of substances according to the kinetics discussed above. Hill8 has made detailed calculations of a special case in the photochemical decomposition of acetone; but, because of mathematical difficulties, he only allowed first order kinetics for the homogeneous termination process in the region near the walls. (1) K. E. Shuler and K. J. Laidler, J . Chcm. Phys., 17, 1212 (1949). (2) N. Rnshevsky, "Mathematical Biophysics," Revised edition, University of Chicago Press. Chicago, Ill., 1948. (3) T. L.Hill, J . Chcm. Phys., 17, 1125 (1949).

While we were considering the interpretation of data on some gas-phase chlorination reactions, we developed some expressions which are of general applicability to photochemical reactions of this type. This treatment should directly apply to some systems, and a t the very least it will provide criteria to determine whether more complicated calculations are necessary. NOMENCLATURE c = local concentration of atoms or radicals per ml. k = rate constant for second-order homogeneous termination of chains in (atom/ml.)-' set.-' (This is twice the customary rate constant based on number of collisions) r = distance from center of reaction vessel in cm. t = cr v = ( B R T / s M ) l / : = average velocity of atom or radical 111 cm./sec. D = diffusion coefficient of atom or radical in cm.Z/sec. Q = rate of production of chains in atom/ml. sec. (This may be as great as twice the rate of absorption of radiation in quanta/ml. sec.) CY = recombination coefficient or probability that an atom or radical striking the wall will undergo recombination B = fraction of chains terminating on the wall y = r / p = (k/Q)%c = ratio of local concentration t o concentration which would prevail in absence of wall reaction p = o/4(Da & ) ' I d p = (Qk/D$L/w c = p - 7 7 = (k8/D2Q)'/at (4) L. Fowler and J. J. Beaver, THISJ O U R N A L , 73 i n M. Noyes and I.. Fowler, i b i d . , 73, 3043 (1931).

~JIXSS;

K.

RICHARD M. NOYES A subscript 0 refers to a value a t the center of the reaction vessel. A subscript w refers t o a value a t or immediately adjacent t o the wall of the reaction vessel.

Mathematical Development Let us postulate a spherical flask throughout which radiation is being absorbed uniformly.5 Let the rate of homogeneous production of atoms or radicals be Q atoms/ml. sec. If the absorption of each quantum produces two atoms i i i the primary process, as in the photodissociation of a halogen, then Q will be twice the rate of absorption of radiation in quanta/ml. sec. Finally, let the rate of disappearance of atoms by homogeneous recombination be k c 2 , where c is the concentration of atoms per ml. and k is the rate constant for the disappearance of atoms. Since two atoms are removed in each effective collision, k is twice the rate constant which is usually calculated for such a proccss.

In most photochemical systems the concentrations of active intermediates are so small that they may be considered to be in a stationary state. Then by the use of Fick's sccond ldw we obtain

+

DV'C Q - kCZ 0 (1) In this expression D is the diffusion coefficient of atoms in cin.2/sec. Strictly speaking, D is a function of concentration and hence of position in the reaction vessel; however, the concentration of atoms is so small that changes do not affect diffusion properties of the medium, and secondary chemical reactions are usually so slow compared to diffusion that there are no significant gradients in any other concentrations. Therefore, we shall lose little in generality if we assume D to be constant throughout the system. Since the system is spherically symmetrical, we can rewrite Equation (1) as

Vol. 73

equations presented in the Nomenclature section. It is apparent that a t different positions in any specific reaction vessel values of r , p and y vary directly as t, I and c, respectively. In terms of the new variables Equation (3) may be rewritten as __ d2T = r" - P (4) dP3 P For the situation of interest this equation is satisfied by a family of curves the members of which can be identified by a single parameter. We have chosen to use the parameter yo, the value of y when p = 0. It is apparent that 70 is the ratio of the concentration of atoms at the center of the flask to the concentration which would exist everywhere if there were no recombination on the walls. Solutions of Equation ( 4 ) for various values of 70are shown in Fig. 1; the details of the solution are described in a Mathematical Appendix. The physical significance of the solutions is more apparent if y is plotted against p . Curves of this sort are shown in Fig. 2. To a good f i s t approximation for y > 0.9 or for p < 2 these curves can be described analytically by the expression&

+

1 - YO ___-skh(1 YO)'/PP (5) (1 YO)'/PP For situations other than small p or large y,different approximations are more satisfactory. = 1-

+

where I is the distance of a point from the center of the reaction vessel. In order to solve this equation we have found it simplest to make the substitution t = cr. Equation (2) then becomes d2t

= 0 (3) dr2 Dr For the sake of generality in the solution, it is convenient 10 define the dirncnsionless variables r , p and y by incans of thc

I. 0

,

\ a

\

,

2

6

4

8

10

P

I'ig. d.-L)epeiidence of concentration ( y ) on radius ( p ) . Curves correspond to values of 1 - yo of 0.9, 0.7, 0.5, 0.3, 0.1, 0.01, 10-3, 10-5 and 10-6.

Discussion

0

2

4

f

S

10

P

Fig. l.---l)ependeiice of T and p . Curves correspoiid of I - y o of o 7 , o 5 , o 3 , 0 1, o 01,10-3,10-4,io-~, I 0-6 lllld 10 -7.

to

(.j)Of

course photocheniical reaction vessels are usually not spher-

i w l , arid the absorption of radiation in a system of finite thickness is iirver absolutely uniform. I n later sections m e shall cotisidcr modificatitin:, for real systems.

Maximum Influence of Wall.-Let pw be the value of p a t the wall of the reaction vessel. If the wall is so efficient at removing atoms that the concentration in its immediate neighborhood is negligibly small, the system may be represented by the curve of the type shown in Fig. 2 for which y vanishes a t pw. If pw is large, the curve is very steep, and no matter how efficient the walls are they may reduce the Concentration of atoms in only a very localized portion of the vessel. For p w > 5 the curves are nearly exponential and may be approximated by an expression of the form = ]

-

f-l.K(P\"

-

p)

(6)

Then y will be 1 - l / e when pw - p = 0.9, and we riiay approximate the situation in a photochemical system by saying that the concentration

-

(6) If t h e homogeneous chain termination is first order in atoms so i h . r t llie r a t e i i IC'(, an exact solution of the diffusion equation is y' I - 1 1 - > 'oi,;iiiti p' where - , I = ( f i ' / @ ) c and p t = ( k ' / ~ ) ' / z r ,

July, 1951

WALL EFFECTSI N PHOTOCHEMICALLY

INDUCED CHAIN

3041

REACTIONS

of atoms is (Q/k)'/z in most parts of the system efficiencies very much less than unity. Let us but is zero in a shell which goes in from the wall a define a as the recombination coefficient or probdistance of 0.9(D2/QK)'I4cm. This distance is ability that an atom which strikes the wall will not about two-thirds of the root mean square distance re-enter the gas phase as an atom. If the heteroan atom diffuses during its mean lifetime, the quan- geneous kinetics are first order in atoms, a will tity which Hilla proposes as the criterion for the be independent of cw, the concentration in the gas distance through which the walls exert an influence. next to the wall; however, the argument is not Either criterion would seem to be a satisfactory affected even if a is allowed to be a function of cw. way to calculate whether the walls can exert any It may also be that some atoms are adsorbed on the wall but subsequently re-evaporate as atoms, Of significant influence on a reaction. Although the curves in Fig. 2 were calculated course in the steady state we cannot distinguish for a spherical flask, the conclusions of the pre- between this effect and simple elastic collisions. ceding paragraph are applicable t o any vessel By combining Fick's f i s t law with the standard provided 0.9(D2/Qk)' 1 4 is small compared to the formula for rate of collisions with a surface, we can linear dimensions. Similarly, the value of Q write -(d~/dr)w = C ~ V C Y / ~ D (7) can be allowed to vary in different parts of the flask to take account of changes throughout the where v is the average velocity of an atom in em./ system. However, the approximations are suf- sec. Then by appropriate substitution we obtain ficiently large that such corrections cannot be justified unless the greater part of the incident radiation is absorbed in the reaction vessel. If p and pw are known, Equation (8) may be It is instructive to consider the magnitudes of the quantities which may be involved. I n many used to obtain a relationship between yw and a photochemical experiments Q is of the order of and thus to calculate the effect which a given re10l2 or more atom/ml. sec. If homogeneous re- combination coefficient will have on the concencombination takes place a t every collision between tration of atoms adjacent to the wall. The relationship is not simple for small values of pw, but (atom/ml.)-' atoms, k will be of the order of sec.-l. At pressures of one atmosphere D is about for p w > 5 we may use the approximation of Equa0.2 cma2/sec. Then we find that 0.9(D2/Qk)'/'is tion (6) and obtain of the order of a few tenths of a centimeter, and the dy/dp = -1.15(1 - y ) (9) walls may influence a significant but calculable and yw = l . l 5 / ( p a f 1.15) (10) fraction of the volume of the system. On the other hand, homogeneous recombination of atoms I n the general case in which a is a function of and perhaps of many radicals takes place much less yw, little has been gained by the derivation of frequently than on every collision. It therefore Equation (10). However, many heterogeneous appears that wall effects in gas-phase reactions can recombinations appear to exhibit first order kinetics be eliminated with certainty only with very high so that a is independent of yw. We have illuslight intensities and/or the use of total gas pres- trated these systems in Fig. 3, where yw from sures considerably in excess of atmospheric. Equation (10) is plotted against log p a . We see These criteria are seldom obeyed. that if p a > l l 5 , the concentration of atoms near The situation in liquid phase is very different. the wall has fallen to less than 1% of that in the Here Q may still be of the order of 10l2 atoms/ml. main body of the vessel; while if pa